by 10iii
AIR is a toolkit that compresses tool output, such as file reads, command output and test results, before it reaches an AI model's context window. It is available as a CLI, a library and plugins.
PluginContact for pricing
Step 1: Install Node.js 18 or later.
Step 2: Install the CLI with npm install -g @10iii/air.
Step 3: Run air read with the skeleton option to view a file's signatures.
Step 4: Run air bash, air grep or air test to compress command, search or test output.
Step 5: Or import the compressors from @10iii/air-core in your own code.
Step 6: For OpenCode, add @10iii/air-oc-plugin to the plugin list in opencode.json.
No. The README says compression is rule-based and needs no external API keys.
The README gives typical token savings by compressor, for example 90% or more for test output and 60 to 90% for command output.
The README says anonymous usage statistics are collected, including content hashes and compression ratios but not the content. You can change this with air config.
MIT.
Requirements
CtxLint MCPctxlint checks AI agent context files and MCP server configs against your real codebase and flags drift before your agent acts on it. It runs as a CLI, CI step, pre-commit hook or MCP server.
Akash NetworkAkash Network is a decentralized marketplace where independent providers bid to host your containers and GPU workloads. You describe what you need and pick a bid.